Skip to content

lorinescu/dubliner

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

77 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Simple screen scraper for Elite:Dangerous Power Play and system data.

Install JRE 8u60 or higher.

Download latest release from here , unpack and click on start.bat .

To use from command line, open a command prompt, switch to the Dubliner directory and type:

 start.bat -cli some\image\file-or-directory

Dev environment:

  • Source code sits in bitbucket
  • If you get a "Error: Illegal min or max specification!" error set in your environment "LC_ALL=C" (known bug)
  • Running the app or its tests will fail unless the "dataPath" setting (in conf/settings.json ) points at the training data dirs in dubliner-data repository or your own trained data. In my environment I check out dubliner and dubliner-data in the same parent directory, switch to dubliner directory and create a symbolic link to the data repo (in linux: ln -s ../dubliner-data data or in windows: mklink /D data ..\dubliner-data).
  • JavaFX SceneBuilder might not load some fxml file. Usually this is related to having custom/3rd party UI controls (ex: ControlFX) that are not on the classpath. The solution is to update the classpath by editing app/SceneBuilder.cfg file in the SceneBuilder installation directory.

Notes:

Changelog:

  • 0.0.4
    • improved power name character recognition - Tesseract does not deal well with huge fonts so scaling up the power name rectangle does not help at all
    • additional corrections for marker words variations
    • configurable csv export separator
    • missing separator in csv header
    • WIP - mock-up UI for system reports
    • WIP - ImageApi plumbing changes to support other type of screenshots besides PP
  • 0.0.3
    • support for Preparation and Expansion tabs
    • reduce memory footprint - all original screenshots and the extracted segments no longer sit in memory and are loaded only when needed
    • notification sound after finishing an OCR task
    • when clicking on an a text field with the OCR results show a popup with raw OCR output
    • show more information about data extraction progress: current processed image, % of work done
    • use populated system list from EDDB as a whitelist for system name corrections
  • 0.0.2
    • control screen data extraction for all powers
    • using OpenCV for automatically identifying data rectangles
    • -cli parameter now accepts directories and will deep search for bmp files
    • -debug parameter will generate excessive debugging data in out/ dir
    • removed some unused UI elements and improved existing
    • updated system name corrections
  • 0.0.1 - initial version

TODO:

  • (request) dynamically reload settings and allow to manually re-trigger an individual image processing
  • (request) traffic report, bounty hounter report, crime report and top 5 bounties
  • system names have a common OCR problem - O is recognized as 0 almost every time
  • in a few cases the undermine trigger value is not extracted correctly as no space is "seen" between "TRIGGER" and the actual counter
  • Control and Expansion screen data extraction flows are similar, they could be merged somehow
  • scale down a bit the layout, on smaller screens not everything is visible
  • in the system ListView mark entries that have invalid data
  • use SuperCSV or similar for generating CSV files instead of joining strings with commas
  • have a (user) configurable worker pool for processing images in parallel
  • to keep the game's screenshot directory clean, move interesting images to a spool directory and process them there
  • installer, see NSIS

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages